Old Systems, New Tools
There is a difference between preserving an old idea and refusing to move forward.
I do not work with xBase ideas because every old design was good. Many were limited by memory, storage, operating systems, and assumptions that no longer apply. I return to them because some of those systems were unusually direct.
A table was visible. A record pointer had meaning. Commands were readable. A user could inspect the work rather than merely trust a hidden service.
Carry the useful parts forward⌗
Modern tools give us better compilers, testing, version control, continuous builds, portable libraries, graphical interfaces, and far more computing power. They also make it easy to hide a modest problem beneath a tall stack of abstractions.
The useful course is neither nostalgia nor fashionable replacement. It is selection.
Keep what was clear:
- understandable data structures
- explicit commands
- inspectable state
- durable files
- repeatable scripts
- documentation close to the work
Replace what was limiting:
- narrow sizes and address spaces
- fragile global assumptions
- platform dependence
- untested behavior
- undocumented side effects
- artificial format ceilings
The long view⌗
A good modernization project should answer two questions.
First: what can the new tools do that the old system could not?
Second: what did the old system make obvious that the new one is in danger of concealing?
The best result respects both answers.